| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343 |
- package ipblocklist
-
- import (
- "bufio"
- "context"
- "fmt"
- "io"
- "io/ioutil"
- "net"
- "net/http"
- "net/url"
- "os"
- "regexp"
- "strings"
- "sync"
- "time"
-
- "github.com/9seconds/mtg/v2/mtglib"
- "github.com/kentik/patricia"
- "github.com/kentik/patricia/bool_tree"
- "github.com/panjf2000/ants/v2"
- )
-
- const (
- fireholIPv4DefaultCIDR = 32
- fireholIPv6DefaultCIDR = 128
- )
-
- var fireholRegexpComment = regexp.MustCompile(`\s*#.*?$`)
-
- type Firehol struct {
- ctx context.Context
- ctxCancel context.CancelFunc
- logger mtglib.Logger
-
- rwMutex sync.RWMutex
-
- remoteURLs []string
- localFiles []string
-
- httpClient *http.Client
- workerPool *ants.Pool
-
- treeV4 *bool_tree.TreeV4
- treeV6 *bool_tree.TreeV6
- }
-
- func (f *Firehol) Contains(ip net.IP) bool {
- if ip == nil {
- return true
- }
-
- ip4 := ip.To4()
-
- f.rwMutex.RLock()
- defer f.rwMutex.RUnlock()
-
- if ip4 != nil {
- return f.containsIPv4(ip4)
- }
-
- return f.containsIPv6(ip.To16())
- }
-
- func (f *Firehol) containsIPv4(addr net.IP) bool {
- ip := patricia.NewIPv4AddressFromBytes(addr, 32)
-
- if ok, _, err := f.treeV4.FindDeepestTag(ip); ok && err == nil {
- return true
- }
-
- return false
- }
-
- func (f *Firehol) containsIPv6(addr net.IP) bool {
- ip := patricia.NewIPv6Address(addr, 128)
-
- if ok, _, err := f.treeV6.FindDeepestTag(ip); ok && err == nil {
- return true
- }
-
- return false
- }
-
- func (f *Firehol) Run(updateEach time.Duration) {
- ticker := time.NewTicker(updateEach)
-
- defer func() {
- ticker.Stop()
-
- select {
- case <-ticker.C:
- default:
- }
- }()
-
- if err := f.update(); err != nil {
- f.logger.WarningError("cannot update blocklist", err)
- }
-
- for {
- select {
- case <-f.ctx.Done():
- return
- case <-ticker.C:
- if err := f.update(); err != nil {
- f.logger.WarningError("cannot update blocklist", err)
- }
- }
- }
- }
-
- func (f *Firehol) Shutdown() {
- f.ctxCancel()
- }
-
- func (f *Firehol) update() error { // nolint: funlen, cyclop
- ctx, cancel := context.WithCancel(f.ctx)
- defer cancel()
-
- wg := &sync.WaitGroup{}
- wg.Add(len(f.remoteURLs) + len(f.localFiles))
-
- treeMutex := &sync.Mutex{}
- v4tree := bool_tree.NewTreeV4()
- v6tree := bool_tree.NewTreeV6()
-
- errorChan := make(chan error, 1)
- defer close(errorChan)
-
- for _, v := range f.localFiles {
- go func(filename string) {
- defer wg.Done()
-
- if err := f.updateLocalFile(ctx, filename, treeMutex, v4tree, v6tree); err != nil {
- cancel()
- f.logger.BindStr("filename", filename).WarningError("cannot update", err)
-
- select {
- case errorChan <- err:
- default:
- }
- }
- }(v)
- }
-
- for _, v := range f.remoteURLs {
- value := v
-
- f.workerPool.Submit(func() { // nolint: errcheck
- defer wg.Done()
-
- if err := f.updateRemoteURL(ctx, value, treeMutex, v4tree, v6tree); err != nil {
- cancel()
- f.logger.BindStr("url", value).WarningError("cannot update", err)
-
- select {
- case errorChan <- err:
- default:
- }
- }
- })
- }
-
- wg.Wait()
-
- select {
- case err := <-errorChan:
- return fmt.Errorf("cannot update trees: %w", err)
- default:
- }
-
- f.rwMutex.Lock()
- defer f.rwMutex.Unlock()
-
- f.treeV4 = v4tree
- f.treeV6 = v6tree
-
- return nil
- }
-
- func (f *Firehol) updateLocalFile(ctx context.Context, filename string,
- mutex sync.Locker,
- v4tree *bool_tree.TreeV4, v6tree *bool_tree.TreeV6) error {
- filefp, err := os.Open(filename)
- if err != nil {
- return fmt.Errorf("cannot open file: %w", err)
- }
-
- go func(ctx context.Context, closer io.Closer) {
- <-ctx.Done()
- closer.Close()
- }(ctx, filefp)
-
- defer filefp.Close()
-
- return f.updateTrees(mutex, filefp, v4tree, v6tree)
- }
-
- func (f *Firehol) updateRemoteURL(ctx context.Context, url string,
- mutex sync.Locker,
- v4tree *bool_tree.TreeV4, v6tree *bool_tree.TreeV6) error {
- req, err := http.NewRequestWithContext(ctx, http.MethodGet, url, nil)
- if err != nil {
- return fmt.Errorf("cannot build a request: %w", err)
- }
-
- resp, err := f.httpClient.Do(req) // nolint: bodyclose
- if err != nil {
- return fmt.Errorf("cannot request a remote URL %s: %w", url, err)
- }
-
- go func(ctx context.Context, closer io.Closer) {
- <-ctx.Done()
- closer.Close()
- }(ctx, resp.Body)
-
- defer func(rc io.ReadCloser) {
- io.Copy(ioutil.Discard, rc) // nolint: errcheck
- rc.Close()
- }(resp.Body)
-
- return f.updateTrees(mutex, resp.Body, v4tree, v6tree)
- }
-
- func (f *Firehol) updateTrees(mutex sync.Locker,
- reader io.Reader,
- v4tree *bool_tree.TreeV4,
- v6tree *bool_tree.TreeV6) error {
- scanner := bufio.NewScanner(reader)
-
- for scanner.Scan() {
- text := scanner.Text()
- text = fireholRegexpComment.ReplaceAllLiteralString(text, "")
- text = strings.TrimSpace(text)
-
- if text == "" {
- continue
- }
-
- ip, cidr, err := f.updateParseLine(text)
- if err != nil {
- return fmt.Errorf("cannot parse a line: %w", err)
- }
-
- if err := f.updateAddToTrees(ip, cidr, mutex, v4tree, v6tree); err != nil {
- return fmt.Errorf("cannot add a node to the tree: %w", err)
- }
- }
-
- if scanner.Err() != nil {
- return fmt.Errorf("cannot parse a response: %w", scanner.Err())
- }
-
- return nil
- }
-
- func (f *Firehol) updateParseLine(text string) (net.IP, uint, error) {
- _, ipnet, err := net.ParseCIDR(text)
- if err != nil {
- ipaddr := net.ParseIP(text)
- if ipaddr == nil {
- return nil, 0, fmt.Errorf("incorrect ip address %s", text)
- }
-
- ip4 := ipaddr.To4()
- if ip4 != nil {
- return ip4, fireholIPv4DefaultCIDR, nil
- }
-
- return ipaddr.To16(), fireholIPv6DefaultCIDR, nil
- }
-
- ones, _ := ipnet.Mask.Size()
-
- return ipnet.IP, uint(ones), nil
- }
-
- func (f *Firehol) updateAddToTrees(ip net.IP, cidr uint,
- mutex sync.Locker,
- v4tree *bool_tree.TreeV4, v6tree *bool_tree.TreeV6) error {
- mutex.Lock()
- defer mutex.Unlock()
-
- if ip.To4() != nil {
- addr := patricia.NewIPv4AddressFromBytes(ip, cidr)
-
- if _, _, err := v4tree.Set(addr, true); err != nil {
- return err // nolint: wrapcheck
- }
- } else {
- addr := patricia.NewIPv6Address(ip, cidr)
-
- if _, _, err := v6tree.Set(addr, true); err != nil {
- return err // nolint: wrapcheck
- }
- }
-
- return nil
- }
-
- func NewFirehol(logger mtglib.Logger, network mtglib.Network,
- downloadConcurrency uint,
- remoteURLs []string,
- localFiles []string) (*Firehol, error) {
- for _, v := range remoteURLs {
- parsed, err := url.Parse(v)
- if err != nil {
- return nil, fmt.Errorf("incorrect url %s: %w", v, err)
- }
-
- switch parsed.Scheme {
- case "http", "https":
- default:
- return nil, fmt.Errorf("unsupported url %s", v)
- }
- }
-
- for _, v := range localFiles {
- if stat, err := os.Stat(v); os.IsNotExist(err) || stat.IsDir() || stat.Mode().Perm()&0o400 == 0 {
- return nil, fmt.Errorf("%s is not a readable file", v)
- }
- }
-
- if downloadConcurrency == 0 {
- downloadConcurrency = 1
- }
-
- workerPool, _ := ants.NewPool(int(downloadConcurrency))
- ctx, cancel := context.WithCancel(context.Background())
-
- return &Firehol{
- ctx: ctx,
- ctxCancel: cancel,
- logger: logger.Named("firehol"),
- httpClient: network.MakeHTTPClient(nil),
- treeV4: bool_tree.NewTreeV4(),
- treeV6: bool_tree.NewTreeV6(),
- workerPool: workerPool,
- remoteURLs: remoteURLs,
- localFiles: localFiles,
- }, nil
- }
|